Symmetry-Breaking Convergence Analysis of Certain Two-layered Neural Networks with ReLU nonlinearity

Authors: Yuandong Tian

ICLR 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Simulations verify our theoretical analysis. ... Sec. 5 shows that simulation results are consistent with theoretical analysis.
Researcher Affiliation Industry Yuandong Tian Facebook AI Research yuandong@fb.com
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any statement or link for open-sourcing its code.
Open Datasets No The paper assumes that the input x follows Gaussian distribution (synthetic data assumption) but does not mention the use of any publicly available or open real-world dataset with access information. It states: "We assume that the input x follow Gaussian distribution." and "We prepare the input data X with standard Gaussian distribution".
Dataset Splits No The paper does not mention specific training, validation, or test dataset splits for any real-world data. It analyzes theoretical dynamics with assumed Gaussian input distribution.
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory) used for running its simulations.
Software Dependencies No The paper does not specify any software dependencies with version numbers used for its simulations or analysis.
Experiment Setup No The paper describes the theoretical setup and assumptions (e.g., Re LU nonlinearity, Gaussian input, teacher-student setting) but does not provide specific experimental setup details such as hyperparameters (learning rate, batch size, epochs, optimizers) for training a neural network model. It focuses on the dynamics analysis.